Background: Avaailable data sugget that if energy expenditure (EE) is involved in the etiology of obesity, the component of interest is activity-related EE (AEE). However, most prior studies examining the role of EE did not ensure metabolically stable conditions, include controls, or distinguish subjects according to ethnicity or predisposition to obesity. Building on our previous studies, we propose to use a carefully controlled model of the post-obese state by reducing obese black and white women to a norml-weight post-obese state, and by comparing them with never-obese controls. In these groups we will test the hpotheses that 1) reduced spontaneous AEE is a causative factor, rather than an effect, of the obese state, and that 2) the cause of reduced AEE is an underlying disorder of muscle substrate oxidation which, in turn, causes increased physiologic and perceived exercise difficulty in obese and post-obese persons, compared to never-obese controls. Objectives: To compare spontaneous sedentary AEE (by respiration chamber), free-living AEE (doubly labeled water), exercise difficulty (by standardized exercise tests), exercising muscle oxidative capacity (by magnetic resonance spectroscopy), and body composition (4-compartment model) in African-American and Caucasian obese and post-obese women, relative to never-obese controls. Design: Utilizing a General Clinical Research Center, 50 moderately obese, normoglycemic, premenopausal black and white women with a familial predisposition to obesity will be studied in the obese state and after reduction to a normal-weight, post-obese state, using a low-calorie diet. The post-obese will be group-matched on age, ethnicity, and body composition with 50 never-obese controls without a family history of obesity. Using this design, we will determine 1) if, in the post-obese state, AEE, exercise difficulty, and muscle energy metabolism are abnormal relative to never-obese controls (a difference suggesting a causative role in the etiology of obesity), and 2) if the obese state per se is an additional cause of abnormalities in these study parameters.